Search papers, labs, and topics across Lattice.
This paper introduces CoCo, a coherence-aware co-optimization framework that integrates task mapping and routing in many-core systems by employing a unified cost model that accounts for communication costs, coherence overhead, and load imbalance. The approach addresses the significant overhead introduced by cache coherence, which is often overlooked in existing methods, leading to improved performance metrics. Experimental results demonstrate that CoCo achieves substantial reductions in link utilization, packet delay, and execution time compared to traditional approaches, underscoring the critical role of cache coherence in optimizing many-core system performance.
CoCo achieves an 88.46% reduction in link utilization while effectively managing coherence overhead, revolutionizing many-core system optimization.
The rapid growth of data-intensive applications increases communication demands in many-core systems, where cache coherence, while essential for correct communication and data consistency, introduces substantial overhead due to frequent data sharing and coherence activities. As system scale and workload complexity grow, the resulting coherence traffic intensifies communication pressure, making the co-optimization of task mapping and routing essential for improving system performance. However, most existing approaches overlook cache coherence, leaving a substantial portion of coherence-induced communication unaccounted for and creating a mismatch between optimization objectives and actual communication patterns. Furthermore, by employing separate cost evaluators for mapping and routing, these approaches complicate objective coordination, may lead to conflicting decisions, and fail to capture the coherence-induced coupling between the two stages. To address these challenges, we propose CoCo, a coherence-aware co-optimization framework that jointly integrates task mapping and routing under a unified cost model for realistic scenarios. This unified model integrates communication cost, coherence overhead, and load imbalance into a single objective, enabling coherence-aware decision-making and effective trade-offs among optimization goals. Guided by this model, CoCo combines coherence-guided task mapping with reinforcement learning-based routing, where directional link weights are adjusted according to communication behavior to improve traffic distribution, enabling coherence-aware co-optimization for many-core systems. Experimental results show that CoCo reduces link utilization by 88.46%, packet delay by 17.40%, and execution time by 17.58% compared with existing approaches, highlighting the importance of cache coherence in co-optimization design.
